Open PL LLM - PPC (multiple choice, 5-shot): leaderboard
Polish Paraphrase Corpus: classifying sentence pairs as exact paraphrases, close in meaning or unrelated. Open PL LLM Leaderboard (SpeakLeash), 5-shot, multiple choice variant: the options are ranked by log-likelihood; accuracy (%), higher is better. Runs not available on the Hugging Face Hub are excluded.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around February 2027. 324 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 32B | 83.9 | #544 |
| 2 | QwQ 32B-Preview | 83.2 | #671 |
| 3 | Qwen 2.5 32B Instruct | 82.9 | #509 |
| 4 | Qwen 2.5 72B | 82.9 | #512 |
| 5 | Qwen 2.5 72B Instruct | 82.5 | #469 |
| 6 | QwQ-32B | 82.4 | #422 |
| 7 | Llama 3.1 405B Instruct FP8 | 81.9 | #400 |
| 8 | Qwen 2 72B | 81.9 | #515 |
| 9 | Qwen 2 72B Instruct | 81.8 | #579 |
| 10 | Qwen 2.5 14B | 81.6 | #716 |
| 11 | cogito-v1-preview-qwen-14B | 81.1 | |
| 12 | Qwen 3 32B | 81 | #449 |
| 13 | Qwen 3 30B A3B | 80.9 | #496 |
| 14 | Llama 3.3 70B Instruct | 80.7 | #541 |
| 15 | Qwen 3 14B | 80.3 | #542 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-pl-llm-ppc-multiple-choice-5-shot · How It Works · Data refreshed daily, snapshot 2026-11-05.